envappconfig

envappconfig is intended to provide simple configuration via environment variables, in the same spirit as argparse, which can be useful when developing and deploying 12-factor apps.

Features:

• Autogenerates usage output if an environment variable is missing
• Default settings for missing environment variables
• Functions that transform the environment variable string to the value type you need
• Environment variable prefixes

Basic example

from envappconfig import EnvAppConfig

env = EnvAppConfig(description='Amazing app')
env.add_env('port', default=1234, transform=int, help='The listen port')
env.add_env('mirror', help='The URL to mirror')
config = env.configure()

# Returns PORT from os.environ transformed to an int,
# or 1234 if PORT does not exist.
config.port

# Returns MIRROR from os.environ,
# or displays usage at env.configure()
# if MIRROR does not exist, then exits.
config.mirror

Adding a prefix

If all the environment variables for the app have the same prefix, it can be specified with the prefix parameter.

from envappconfig import EnvAppConfig

env = EnvAppConfig(prefix='MYAPP', description='Amazing app')
env.add_env('port', default=1234, transform=int, help='The listen port')
env.add_env('mirror', help='The URL to mirror')
config = env.configure()

# Returns MYAPP_PORT from os.environ transformed to an int,
# or 1234 if MYAPP_PORT does not exist.
config.port

# Returns MYAPP_MIRROR from os.environ,
# or displays usage at env.configure()
# if MYAPP_MIRROR does not exist, then exits.
config.mirror

Custom transforms

The transform parameter can be used to specify normal transforms, like int or float (the default is str), but it can also take custom transform functions. The transform function must take a single parameter, which will be filled in with the string value from the environment variable.

env = EnvAppConfig(description='Amazing app')

# Double the timeout specified in the TIMEOUT environment variable,
# or default to 60.
env.add_env('timeout', default=60, transform=lambda x: int(x) * 2, help='Timeout in seconds')
...

Adding more config values

Additional config values can be added to an existing namespace, which can be helpful when there's a config value that needs to be calculated based on other config values.

from envappconfig import EnvAppConfig

env = EnvAppConfig(description='Amazing app')
env.add_env('bind', help='IP address to bind to')
env.add_env('port', default=1234, transform=int, help='The listen port')
config = env.configure()
config.listen = f'{config.bind}:{config.port}'

# Returns the combined bind:port string.
config.listen

Forcing usage display

The environment variable usage output for an app can be forced by defining the ENVAPPCONFIG_SHOW_USAGE environment variable to any value. This is basically like --help for apps that take command line parameters. Forcing usage display can be useful when you've got a pre-built container for an app that uses envappconfig, and you need to see all the environment variable options. For example:

docker run -it --rm -e ENVAPPCONFIG_SHOW_USAGE=1 amazing:latest

or in Kubernetes:

env:
  - name: ENVAPPCONFIG_SHOW_USAGE
    value: "1"

Command Line

There are a couple options for using envappconfig at the command line (eg. when testing).

Prefix

If you've only got a couple environment variables to set, just put them before the command:

PORT=9999 NAME=foo python3 script_using_envappconfig.py

dotenv

If you have more environment variables to set, consider using dotenv. First put your environment variables in a file named .env:

PORT=9999
NAME=foo

Then call dotenv as follows, which will load up the variables from .env for this command:

dotenv run -- python3 script_using_envappconfig.py

You can install the dotenv command line tool with:

python3 -m pip install "python-dotenv[cli]"